INDEX

Explanations

terms related to news and events, specifically focusing on legal cases and political developments

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agus

-0.70

ASE

-0.65

UGE

-0.62

Â£

-0.62

liest

-0.61

adden

-0.61

winter

-0.60

 Rampage

-0.59

endi

-0.59

click

-0.58

POSITIVE LOGITS

 includes

0.86

 included

0.84

 noted

0.80

 referred

0.78

 noteworthy

0.78

 contributed

0.78

 served

0.77

 cautioned

0.75

 Accessed

0.74

 optionally

0.74

Activations Density 1.257%

terms related to news and events, specifically focusing on legal cases and political developments

No Comments

No Known Activations